Reinforcement Learning of Intelligent Characters in Fighting Action Games
暂无分享,去创建一个
In this paper, we investigate reinforcement learning (RL) of intelligent characters, based on neural network technology, for fighting action games. RL can be either on-policy or off-policy. We apply both schemes to tabula rasa learning and adaptation. The experimental results show that (1) in tabula rasa leaning, off-policy RL outperforms on-policy RL, but (2) in adaptation, on-policy RL outperforms off-policy RL.
[1] Sung Hoon Jung,et al. Exploiting Intelligence in Fighting Action Games Using Neural Networks , 2006, IEICE Trans. Inf. Syst..
[2] Sung Hoon Jung,et al. Adaptation of Intelligent Characters to Changes of Game Environments , 2005, CIS.
[3] K. R. Dixon,et al. Incorporating Prior Knowledge and Previously Learned Information into Reinforcement Learning Agents , 2000 .
[4] R. Lippmann,et al. An introduction to computing with neural nets , 1987, IEEE ASSP Magazine.